INDEX

Explanations

adjectives and adverbs indicating intensity or degree

oai_token-act-pair · gpt-3.5-turbo

expressions indicating ongoing relevance or popularity

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

attery

-0.68

assi

-0.67

lez

-0.61

imov

-0.61

erity

-0.61

 Mortgage

-0.61

Tesla

-0.60

 Advice

-0.59

andum

-0.59

Interstitial

-0.59

POSITIVE LOGITS

 virgin

0.76

 intact

0.74

 untouched

0.73

 adolesc

0.73

 scratching

0.71

 unsolved

0.71

 unanswered

0.70

birth

0.70

 reeling

0.66

 relevance

0.65

Activations Density 0.276%

adjectives and adverbs indicating intensity or degree

expressions indicating ongoing relevance or popularity

No Comments

No Known Activations